AI Daily · Nov 13, 2026 · FriAI Daily
ICREAT AI NEWS
NVIDIA announces CEO transition: Jensen Huang steps down, CFO Colette Kress becomes interim CEO
NVIDIA's board of directors announced that Jensen Huang has resigned as President, CEO and director, effective immediately, with Executive Vice President and CFO Colette Kress serving as interim CEO while continuing as CFO until a permanent successor is named.
Model releases/updates
Google DeepMind releases Gemini 4 Hydrogen model
Google DeepMind released Gemini 4 Hydrogen, aimed at everyday coding and agentic workflows, scoring 51 points at the default high thinking level to beat Argon's 50, and reaching 54 points at the highest xhigh setting.
Edge0 open-sources 145B and 120B on-device models and inference framework
Edge0 open-sourced two on-device models, Edge0-145B and Edge0-120B, along with a unified inference framework and deployment tools. Edge0-145B has 145 billion parameters, with reported peak memory of 29GB, decoding at 80+ tokens/s, prefill at 1100+ tokens/s, and support for 128K–256K context.
Product releases/updates
Anthropic launches Claude Meetings, letting Claude join video meetings
Anthropic launched the beta of Claude Meetings, letting Claude join video meetings as a visible participant, follow the conversation and produce documents, whiteboards and meeting notes in real time. Teams can ask Claude questions during the meeting by voice or chat, with text answers by default and, with permission, calls to connected tools; meeting records, action items and artifacts are saved on the meetings page in claude.ai.
Claude Opus 5.5 becomes the default model for Pro, Max, Ultra and Team plans
Claude Opus 5.5 became the default model for Claude on the Pro, Max, Ultra and Team plans, and Anthropic recommends it as the everyday workhorse for research, analysis and document production.
Industry news
Timeline of the Nvidia CEO Jensen Huang scandal: board moves to remove him, Congress moves to summon him
TechCrunch compiled a reverse-chronological timeline of developments in the relationship between Nvidia CEO Jensen Huang and alleged Chinese spy Alice Huang: Huang admitted to employees on November 8 that he had been unfaithful to his wife, Nvidia's board launched an independent investigation, and the stock fell more than 5% the next day.
Nvidia after Huang's departure and the direction of China chip controls
SemiAnalysis says Jensen Huang is no longer Nvidia's CEO and that the board is investigating his conduct; the author believes this is tied to more than 1.5GW of restricted Nvidia chips flowing into China, and judges that the US intelligence community is an important force behind his departure.
Gary Marcus cites Wall Street Journal report that Nvidia is about to fire Jensen Huang
Gary Marcus cited a Wall Street Journal report saying Nvidia is about to fire Jensen Huang. He argued that if Huang does leave, whether Nvidia's promised IPO investment in Anthropic will proceed becomes a key question, and called Huang a central figure in the AI industry's circular financing. The original text did not specify the basis of the Wall Street Journal report.
From Denny's dishwasher to AI star: Jensen Huang's rise and resignation
Former Nvidia CEO Jensen Huang resigned on Thursday over alleged improper relationships with a Chinese spy and an Nvidia executive, and the company's valuation fell back to about $4 trillion this week. He co-founded Nvidia in 1993 with Chris Malachowsky and Curtis Priem in a Denny's booth; the company launched the GeForce 256 in 1999, crossed a $1 trillion market cap in May 2023, and reached $5 trillion last October.
The Jensen Huang resignation saga: what does China actually want from him
The Verge podcast Decoder discussed Nvidia CEO Jensen Huang's resignation over his dealings with a woman alleged to be a Chinese spy, with guest Kyle Chan, a fellow at the Brookings Institution's John L. Thornton China Center. Chan argued Beijing more likely wants visibility into and influence over US policymaking than chip design data, since Nvidia's commercial relationship with Megaspeed already existed; he also noted the saga could weaken those advocating relaxed export controls and push Congress toward an investigation.
Tips and opinions
Five ways Jensen Huang's removal by Nvidia's board could hit the AI industry
Nvidia CEO Jensen Huang was removed by the board on Thursday after admitting to extramarital affairs, one of which involved a woman suspected by the US government of being a Chinese spy. The article argues he had been the leading champion of open-weight models, self-regulation of agent safety, AI PCs and data center financing, and these areas now lose their biggest advocate; Nvidia has agreed to acquire Hugging Face for $12.9 billion, with the deal expected to close in the first half of next year.
Jensen Huang steps down as Nvidia CEO: what a leadership change at the AI economy's 'central bank' means
Tomasz Tunguz analyzed the impact of Jensen Huang stepping down as Nvidia CEO on the AI economy: over the past two years Nvidia spent $42 billion on M&A, took $61 billion in cash stakes in 229 startups and 13 public companies, and provided $164.5 billion in leasing and cloud capacity guarantees. The author argues that if new management tightens equity investments, payment terms or lease guarantees, the pace of AI buildout will slow; meanwhile open-source startups will lose their biggest buyer, and the share of custom ASICs could accelerate.
Briefs
- Unemployment Labs report says six Chinese models including DeepSeek and Moonshot carry Blackwell hardware fingerprints in their weightsHacker News AI
- The fallout from Jensen Huang's resignation is far from overThe Verge · AI
- Vercel data: AI models in production have a half-life of about six weeksHacker News AI
- OpenAI research: why frontier models increasingly hide their reasoningThe Decoder
- AINews roundup: Google releases Gemini Decision 1 decision modelLatent Space
- vLLM runs DeepSeek-V4.1-Flash on Vera Rubin NVL72: 387 tokens/s per user, throughput 58x that of GB300 NVL72vLLM Blog
- Amazon Bedrock AgentCore runtime update: sessions start 15x faster, compute scaled up to 4xAWS Machine Learning Blog
- IBM releases LLMON markup language and Granite 4.1 3B/8B preview modelsHugging Face Blog
- Replit Agent adds capabilities for purchasing business supplies and running Google AdsReplit Blog
- vLLM-Omni supports MiniMax H3-Image on NVIDIA Vera Rubin, generating 4K in 0.87 secondsvLLM Blog